Detect scroll activity (when LC is not frontmost)

Mike Bonner bonnmike at gmail.com
Sun Dec 25 12:08:00 EST 2016


Actually, the question boils down to "is the user active."  not "what key
is pressed, was a screen scrolled(using what device) , or what button was
clicked."  I could be wrong. /shrug  The only recent issue that needed to
be solved is that scroll is an activity too, and there was no easy way to
get lc to see that as an action.

This ignores the whole question and lets lc poll the os to see if there was
any HID action at all-- this (if I understand correctly)  being the
ultimate goal.
So yeah, if you scroll with you belkin, its a scroll.  As on a mighty
mouse. But for this purpose (again, if I understand correctly)  it doesn't
matter.  An action of any type, by a user (typing, scrolling, clicking,
using a touch screen) shows that the user is actually sitting at the
computer.

If you have a different needs that just tracking idle/active time, this is
not for you.
Otherwise, there is no longer a need to look at keysdown() or
mousescreenloc at all.  Just grab the number from the os. If the mouse has
moved recently, it'll be small. Same for keypresses, scrolls, clicks,
screen touches.

So as in the example of catatonia.. Was it a poke, a tickle, a prod, a
kick..  If all you're looking for is signs of brain activity, it doesn't
matter.  (simplified of course, because in those cases examining responses
to different stimuli DOES matter during evaluation.. but not so with the
desired goal)

On Sun, Dec 25, 2016 at 9:51 AM, Richmond Mathewson <
richmondmathewson at gmail.com> wrote:

> Aha: thanks.
>
> On 12/25/16 6:39 pm, Mike Bonner wrote:
>
>> The script does an end run of the whole situation.  The os itself is
>> keeping track of the idle time between user events.  All the script does
>> is
>> grab the current value. And since only HID (human interface devices) are
>> tracked, any mouse/keyboard activity in any app of the system will reset
>> the timer.  So technically no, the scroll won't "register" in the lc stack
>> (meaning it won't cause a handler to fire), but the OS does track HID
>> actions..  All the stack does is request the information from the os (in a
>> loop), that information being the time since the last user activity.
>>
>
> So, the inevitable question is how one would use an idle time value to tell
> one that the end-user his performed a scroll (and whether up or down), as
> all
> those idle time values are are times in (?) micro-seconds.
>
> I assume (?) that, somewhere in the belly of the beast (Mac OS) a HID
> performed action
> must register as such, and also as WHICH HID was used, and WHAT action was
> performed on
> that HID.
>
> A forward scroll on my Belkin Nostromo is just as much a forward scroll as
> one on my mighty mouse.
>
> The question that started this thread concerns NOT whether one can pick up
> signals that HIDs are
> being used, but when they ARE being used, which of the activities being
> performed is a scroll.
>
> This seems remarkably like the problem with other people:
>
> 1. We can generally tell when brain activity is going on in other people
> (however, c.f. catatonia),
> and we can stick electrodes into parts of the human brain so that we can
> pick up electric pulses
> that tell us when the brain is receiving signals from outside the body.
>
> 2. What cannot (as far as I am aware) be worked out (if one is not
> cheating and looking at who
> is poking your volunteer in the stomach with a chopstick) is what is being
> done to make the brain
> register those signals.
>
> Richmond.
>
>
>>
>>
>> On Sun, Dec 25, 2016 at 9:29 AM, Richmond Mathewson <
>> richmondmathewson at gmail.com> wrote:
>>
>> Does that mean that if, say, I have a stack running your script in the
>>> stackScript
>>> and I'm scrolling a window in Firefox that that scrolling will register
>>> in
>>> the LC stack?
>>>
>>> The reason I am asking that question is because I don't quite understand
>>> how one effect a mouseUp
>>> while one is scrolling with one's mouse at the same time and the mouseUp
>>> not affecting the frontmost app.
>>>
>>> Richmond.
>>>
>>>
>>> On 12/25/16 5:56 pm, Mike Bonner wrote:
>>>
>>> I have an answer..
>>>>
>>>> Heres a sample script:
>>>> local sRunning
>>>>
>>>> on mouseUp
>>>> if sRunning is empty then put false into sRunning
>>>> put not sRunning into sRunning
>>>> loopit
>>>> end mouseUp
>>>>
>>>> command loopit
>>>> if sRunning then
>>>> put the last word of (shell("ioreg -c IOHIDSystem |grep Idle")) into
>>>> tIdle
>>>> put tIdle / 1000000000 into field 1
>>>> send "loopit" to me in 2 sec
>>>> end if
>>>> end loopit
>>>>
>>>> The script is in a button, and I have a single field on the card.  The
>>>> math
>>>> is done to convert to seconds of idle.
>>>>
>>>> The are only 2 disclaimers here.  First is that the value returned pre
>>>> 10.3
>>>> is hex so you'd have to handle that if you have an earlier osx.  10.3
>>>> and
>>>> after this solution should work fine.
>>>>
>>>> The second issue is is that on mac 10.12, the idle time won't update on
>>>> typing.  Its an osx issue for that specific version, but worst case you
>>>> already have a method to track keypresses.
>>>>
>>>> On Sun, Dec 25, 2016 at 8:21 AM, Paul Dupuis <paul at researchware.com>
>>>> wrote:
>>>>
>>>> On 12/25/2016 10:05 AM, Terry Vogelaar wrote:
>>>>
>>>>> So it starts to become clear that it might not be possible to do what I
>>>>>>
>>>>>> want. Although I hope to be wrong about that.
>>>>>
>>>>> I think it is very unlikely you can do this in LC - without externals
>>>>> or
>>>>> LCB widgets from "infinite Livecode".
>>>>>
>>>>> The active mouse and keyboard drivers capture events from these devices
>>>>> and pass that information to the operating system, which massages the
>>>>> data and passed a higher level of events on to the active application,
>>>>> which looks for such events and handles them. In the case of the
>>>>> LiveCode engine - or any app built on the LC engine - that is executing
>>>>> applicable messages for your scripts to handle.
>>>>>
>>>>> Most productivity tracking software works by effectively inserting code
>>>>> into where the device drivers meet the operating system, so that mouse
>>>>> and keyboard events are captured by the productivity app's as well as
>>>>> being sent by the OS to the active application as normal.
>>>>>
>>>>> Using LCB and LC9.0 you might be able to write an LCB widget that does
>>>>> this, but I am not familiar enough with current OSX APIs for event
>>>>> capture or drivers under OSX to or the state of work in LC9.0 on
>>>>> integrating OS API calls to say for sure.
>>>>>
>>>>> You are unlikely to be able to do what you want in LiveCode script
>>>>> alone.
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> use-livecode mailing list
>>>>> use-livecode at lists.runrev.com
>>>>> Please visit this url to subscribe, unsubscribe and manage your
>>>>> subscription preferences:
>>>>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>>>>
>>>>> _______________________________________________
>>>>>
>>>> use-livecode mailing list
>>>> use-livecode at lists.runrev.com
>>>> Please visit this url to subscribe, unsubscribe and manage your
>>>> subscription preferences:
>>>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>>>
>>>> _______________________________________________
>>> use-livecode mailing list
>>> use-livecode at lists.runrev.com
>>> Please visit this url to subscribe, unsubscribe and manage your
>>> subscription preferences:
>>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>>
>>> _______________________________________________
>> use-livecode mailing list
>> use-livecode at lists.runrev.com
>> Please visit this url to subscribe, unsubscribe and manage your
>> subscription preferences:
>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>
>
> _______________________________________________
> use-livecode mailing list
> use-livecode at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your
> subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-livecode
>



More information about the use-livecode mailing list